The German Uranium Miners Cohort Study - First results

Transcription

1 The German Uranium Miners Cohort Study - First results M.Kreuzer, B.Grosche, A.Brachner, K.Martignoni, M.Schnelzer and W.Burkart Federal Office for Radiation Protection, Institute of Radiation Hygiene, Ingolstaedter Landstr. 1, Neuherberg, Germany, mkreuzer@bfs.de INTRODUCTION In 1994, Lubin et al. (1) published a meta-analysis of eleven miner cohort studies conducted in the USA, Canada, the Czech Republic, Sweden, France, Australia and China. The analysis included more than 60,000 miners and about 2,600 lung cancer deaths. The results have demonstrated that underground miners are at an increased risk of lung cancer due to exposure to the radioactive gas radon ( 222 Rn) and its short-lived decay products. Based on the same pooled data set, Darby et al. (2) reported some evidence of risk associated with cancers other than lung cancer like stomach cancer, liver cancer, and leukaemia. A causal relation to radon exposure, however, is still questionable, since no exposure-response relationships were observed. It is estimated that about 400,000 persons have worked at the German Wismut mining company between 1946 and 1989, most of them underground or in uranium ore processing facilities. Until 1990 more than 5,000 of these workers have been compensated in the former GDR as radiation induced lung cancers, and at least 200 to 300 registrations per year are to be expected in the next 10 years (3). The aim of the German uranium miners study, which includes about 60,000 former Wismut miners, is to further evaluate the lung cancer risk associated with radon and its progeny due to cumulative exposure, exposure rate, duration of exposure, time since last exposure, attained age, and the interactions with smoking and arsenic exposure. Moreover, particular concern will be given to the lung cancer risk associated with low levels of radiation and the risk of extrapulmonary cancers. The cohort is now established and the mortality follow-up has been started. This paper presents first characteristics of the study population and provides information about the feasibility and the first results of exposure assessment and the mortality follow-up. MATERIALS AND METHODS Historical background. Already many centuries ago underground mining activies started in the ore mountains (Erzgebirge) of Saxony in Germany. In the earlier centuries silver mining was performed, while later on other metals such as bismuth, cobalt, nickel and tungsten and in the beginning of the 19 th century uranium were mined in the Schneeberg area (4). In 1946, after the Second World War, the Soviet Union started a systematic search for uranium in Saxony and the former Czechoslowakia. The old silver mines were reopened, and the whole mining area was declared a restricted zone under the supervision of the Soviet secret service. In 1947 the Soviet stock corporation SAG Wismut was established and started uranium mining for the nuclear program for the Soviet Union. Later on mining was expanded to areas in Thuringia and in 1954 the company was renamed into the Soviet-German stock corporation (Sowjetisch-Deutsche Aktien-Gesellschaft) SDAG Wismut. It is estimated that about 400,000 persons have worked at the German Wismut company between 1946 and 1989, producing a total of 220,000 tons of uranium in this time period. Thus the former GDR (German Democratic Republic) was the third largest producer of uranium worldwide. With the German reunification in 1990, mining was discontinued and the company closed. Three different time periods of mining activities can be distinguished (table 1). In the period 1946 to 1954, the so-called wild years, exposure to radiation and dust was at its peak due to dry drilling and the lack of forced ventilation. No radon measurements were carried out. Between 1955 and 1970 the working conditions of the miners improved. Dry drilling was replaced by wet drilling, radon measurements started and due to a more efficient ventilation technique concentrations decreased. After 1970 international radiation protection standards were introduced, individual radon measurements were established and the working conditions were stable at a high level of security. 1

2 Table 1: Working conditions at different time periods of employment for former uranium miners at the SDAG Wismut, Germany, Time period Wild years Transition Consolidation characteristics - Many miners (approximately ) - Many mines (approximately 350) - High radiation exposure (up to 300 WLM/a) 1 - No radiation protection - Dry drilling and natural ventilation only - Reduction of miners from to Reduction of number of mines from 350 to 90 - Reduction of radiation exposure from 300 WLM/a to 2 WLM/a - Introduction of radiation protection - Beginning of radon measurements - Stable number of approximately miners - Stable number of approximately 90 mines - International level of radiation protection (2 WLM/a) - Measurement of radon and its decay products 1 : WLM/a: Working level months per year Selection of subjects. The selection of cohort members was based on two personnel files of the Wismut company, which included 120,000 workers with information on gender, year of first employment, and some very rough information on place of work (underground or milling/processing or surface) and location of the mining facility (Thuringia or Saxony). Due to financial restrictions, a random sample of about 65,000 persons was drawn. It was assumed that during the first years also women have worked for at least some time underground. Therefore, it was decided to select for both women and men a random sample stratified by place of work, area of mining and date of first employment to reflect the different mining conditions at the Wismut company. Inclusion criteria was the beginning of employment between 1946 and For the selected sample, data were extracted from the original pay rolls and medical records according to a fixed data catalogue and coding scheme and were entered into the computer. In order to ensure a high quality of the data detailed plausibility checks were carried out and a subset of the data was extracted and entered twice. A working history for each employee was now available in a concept of job periods, where a new period was defined as either a change in mining facility, type of job or place of work. For each job period the exact date of begin and end was recorded. This enabled us to define an internal control group including employees, who worked at the surface only, while exposed subjects were classified into three exposure categories: 1) underground (for at least 180 days), 2) milling or processing (for at least 180 days and less than 180 days underground) and 3) short time exposed workers (less than 180 days underground or/and milling or processing). When the job histories had been evaluated, it turned out that most of the women had never been exposed (working at the surface only), while only a small number of 167 women had worked underground for some time. Therefore women were excluded from the cohort. In summary, a total of 64,459 subjects had been included in the random sample. Inclusion criteria were defined as (1) male, (2) time period of employment more than 180 days, (3) a date of first employment before 1989 and for exposed subjects a date of first exposure before 1989, (4) date of birth after The last criterion was introduced, because a pilot study on the feasibility of the mortality follow-up had demonstrated that certificates of death were often missing before After the evaluation of the data a total of 5,395 subjects had to be excluded, who did not fulfil the final inclusion criteria for the study (see table 2). The cohort members were stratified into three subcohorts according to the different mining conditions. Subcohort A includes subjects who started to work at the Wismut company between 1946 and 1954, subcohort B those starting to work between 1955 and 1970 and subcohort C those starting after

3 Table 2: Definition of the study population Number of subjects Random sample 64,459 Exclusion of - Women 3,737 - Time period of employment less than 180 days or 858 date of first employment after 1989 (for exposed subjects date of first exposure after 1989) - Date of birth before Study population 59,064 Radiation exposure. Exposure to radon and its progeny was estimated by using a job-exposure matrix (JEM), which had been developed by the Miners Occupational Compensation Board (BBG) in Gera (5). The JEM provides the annual dose values for each calendar year of employment between 1946 and 1989, each place of work, and each type of job. More than 900 different jobs and several mining facilities have been evaluated. Since radon measurements were available only after 1955, radon concentrations for the period 1946 to 1954 were estimated based on measurements from 1955, taking into account ventilation rate, vein space and uranium content. Cumulative radiation exposure is given in Working level months (WLM). A working level (WL) is defined as MeV of potential alpha energy/l air. A working level month equals exposure to 1 WL for 170 hours. Potential confounders. Based on different contents of arsenic in the rocks in Saxony and Thuringia, miners working in Saxony were categorized as highly exposed to arsenic and those working in Thuringia as lowly exposed. The quantification of exposure to arsenic will be much more refined by a further job-exposure-matrix, which is in preparation. In addition, individual information on arsenic exposure is available for the time period from 1985 to 1990 for a subset of cohort members. The above mentioned job-exposure-matrix will also provide exposure values for dust. As for arsenic exposure, individual information on dust exposure can also be obtained for a subset of subjects for the time period 1972 to Active smoking habits were derived from annual medical examination files, which were available since Information on smoking habits was categorized for each calendar year into (1) non-smoker, (2) non-smoker since at least 1 year, (3) occasional smoker, (4) smoker of pipes or cigars only, (5) smoker of cigarettes since less than 5 years or smoking daily less than 10 cigarettes, (6) smoker of cigarettes since more than 5 years or smoking daily more than 10 cigarettes. This annual information does not allow the calculation of cumulative tobacco consumption (pack-years). For the present first description of smoking habits, the highest category of smoking habits over all years for each subject was used. Follow-Up. In 1997, a pilot study was conducted to investigate the feasibility of a mortality follow-up. It included 485 cohort members whose time of employment lay back many years. In this worst case scenario the loss to follow-up was around 20%. The main reason for loss to follow-up was a high percentage of incomplete addresses of the miners. Several strategies have been developed to decrease this loss to follow-up. Now for all cohort members the complete last address, which was registered at the Wismut company, is available. The first mortality follow-up having started early in 1999 will determine the vital status of cohort members as of December 31, Data are collected from registration offices, Registrar's offices, Public Health Administrations, and others. Registrar Offices were used to get information on vital status and Public Health Administrations to get information on cause of death. Information whether a person (1) is still living at this address, (2) has moved to another known address, (3) has moved without any information on the new address or (4) has died, is extracted from registration offices and provides the vital status of the subjects. For subjects being deceased the date and place of death is documented in the registration offices. This type of information enables us to contact the responsible Public Health Administrations for obtaining the cause of death. In addition to the mortality follow-up, an incidence follow-up from 1965 to 1989 is planned by means of a record linkage with the data set of the former German Democratic Republic's National Cancer registry. 3

4 RESULTS Table 3 shows characteristics of the study population. The cohort presently includes a total of 59,064 men. There are slightly more miners in subcohort A (38%) than in subcohorts B and C (both 31%). Overall, 42,216 subjects worked underground, 4,554 in processing or milling facilities and 1,330 were only exposed short time. 10,964 persons have worked at the surface only. Miners of subcohort A were mainly employed in mines located in Saxony with high levels of arsenic, whereas miners of subcohort C mostly worked in Thuringia, where levels of arsenic exposure were low. The mean age at first exposure was 25 years, being somewhat higher in subcohort A. On average, study subjects were employed at the Wismut company for 13 years. The duration of employment, however, varies strongly between subcohorts (16 years in subcohort A as compared to 7 years in subcohort C). Information on smoking habits is available for 36% of the total cohort. This proportion is considerably lower for members of subcohort A (20%). More than 50% of the miners with known smoking habits were heavy smokers of cigarettes. The overall proportion of non-smokers is estimated to be around 26%. Table 3: Characteristics of the study population stratified by subcohorts A subcohorts B C total Total numbers 22, Occupational radiation exposure Exposure category Underground a 16,875 11,453 13,888 42,216 Processing or milling b 1,795 1,486 1,273 4,554 Short-time exposure (< 180 days) ,330 Surface only 3,652 4,988 2,324 10,964 Area of mines (%) Saxony (high levels of arsenic) 17,506 7,565 3,923 28,994 Thuringia (low levels of arsenic) 1,761 9,299 12,260 23,320 Saxony and Thuringia 3,356 1,539 1,855 6,750 Age at first employment (years) Mean ± SD 27.9 ± ± ± ± 8.6 Duration of employment (years) Mean ± SD 16.4 ± ± ± ± 11.0 Smoking habits Information available 20.4 % 33.0 % 63.4 % 36.3 % Smoking status c (%) totals 4,624 6,069 11,565 22,258 non-smoker 1,481 1,736 2,617 5,834 non-smokers since at least 1 year occasional smoker smokers of cigars and pipes only cigarettes < 5 years or < 10 cig/day ,573 2,428 cigarettes 5 years or 10 cig/day 2,213 3,434 6,607 12,254 a underground for at least six months b processing or milling for at least six months, but less than six months underground c smoking status is based on subjects where information on smoking was available 4

5 Preliminary results of the temporal pattern of radon exposure according to the JEM are given in figure 1. It presents the average radon exposure per calendar year for the job getter for three different mining areas: (1) the so-called mine Objekt 09, where most of the miners of Saxony have worked, (2) other mines in Saxony and (3) mines in Thuringia. The diagram shows, that according to the JEM, the highest dose values are not found for the very first years of mining, but rather for the time period from 1953 to 1958, when the mines had been in operation for several years. From the late 1960ies on radon exposure was very low. 350,0 300,0 250,0 WLM/y 200,0 150,0 100,0 Object 09 Others Saxony Thuringia 50,0 0, Figure 1: Exposure from radon daughters by calendar year in Object 09 and other mines of the Wismut company for the job type getter - Preliminary estimates according to the JEM The mean cumulative radon exposure for subjects working underground was 390 WLM. It differed markedly between workers in mines in Saxony (569 WLM) and subjects working in mines in Thuringia (74 WLM) and additionally within subcohorts A, B, and C in Saxony (844, 140 and 10 WLM, respectively) and Thuringia (261, 163 and 12 WLM, respectively). In table 4 very first results of the mortality follow-up are given. For 89% of the total cohort information about results of the vital status based on the last known address of the miners are available up to now. Out of this group, 59% of the subjects were registered in the registration office as still living at this address. For these subjects the follow-up is finished and the vital status is documented as alive. About 20% of the subjects were known to be deceased and the responsible Public Health Office will be contacted in a next step in order to get a copy of the certification of death. Twelve percent of the study subjects have moved to a known new address and the responsible registration office will be contacted to evaluate their vital status. Up to now the proportion of potential loss to follow-up is about 8%. Most of these persons were not found in the registration office, probably because of wrong names, dates of birth or addresses. For these subjects, an extensive search on these variables is planned, which may slightly reduce loss to follow-up. 5

6 Table 4: Preliminary results of the mortality follow-up Status of the search for the vital status in the register offices as of 20 th December 1999 Status % of cohort comment Results not yet available 11.0 % Results available 89.0 % (100.0 %) known to be alive 59.4 % Finished known to be deceased 20.4 % Start of the search for cause of death moved to another known address 12.0 % Contact of the next responsible register office moved to another unknown address 1.1 % potential loss to follow-up moved to other countries 0.1 % potential loss to follow-up not yet to be found 7.0 % potential loss to follow-up DISCUSSION The German cohort study is the largest single cohort study on uranium miners so far. It provides a similar size as the meta-analysis of Lubin et al. (1), yet it is more homogeneous with respect to data collection and estimation of exposure. A wide range of exposure levels and a relatively long duration of employment at the Wismut will contribute to improve present knowledge on the shape of the dose-response relationship of radoninduced cancer. Moreover, the combined effects of radon and smoking, dust, and arsenic will be assessed. Validation of data. For a subset of subjects individual data on radon exposure in the time period from 1970 to 1989 are available and a validation of the job-exposure-matrix with individual data for these years is planned. In order to validate the causes of death obtained from the mortality follow-up, a record linkage will be done with the pathology files of the former central pathology archive of the Wismut company. The pathology research group under the coordination of the German Cancer Research Center is evaluating this archive, which contains, among other specimens, tissue samples and protocols for 28,975 autopsy cases collected from 1957 to 1992 (4,6). Up to now 5,215 malignant lung tumors, as noted on the death certificate or in the autopsy protocol have been identified by three reference pathologists. The present study is limited potentially by a lack of information on smoking habits which is mainly available for miners of subcohort C (67%), while overall this proportion is small (37%). Moreover present smoking data are not detailed enough for an exposure quantification other than ever/never-smoker. Since smoking is an important confounder of lung cancer risk, it will be necessary to look for other possible sources of information on smoking habits. A nested case-control study of lung cancer within the cohort is in preparation. The objective of this study is to gather more detailed information on potential confounders such as smoking or known or suspected carcinogenic jobs the miners have been employed before or after their work in the Wismut. This information will be obtained by the miners next of kin by means of self-administered questionnaires. The first mortality follow-up will be finished early in 2002, so that first overall risk estimates will be available presumably by the end of However, results of the incidence follow-up for the time period 1965 to 1989 are to be expected one year earlier. The overall number of lung cancer deaths is estimated to be about 3,000. Presently, 1,405 study subjects are already known to have died of lung cancer. This information was derived from either occupational compensation agencies or from the pathological files held at the German Cancer Research Center in Heidelberg. Subcohort C will be of particular interest, since risk estimates from low-level radiation exposed workers might be transferred to the lung cancer risk arising from indoor radon exposure in houses. The mean age of the subjects being alive on December 31, 1998 among the three subcohorts A, B and C however, is estimated to be about 72 years, 61 years and 41 years, respectively (7). Thus the number of deaths in subcohort C will be small in the first mortality follow-up and further mortality follow-ups are envisaged. In conclusion, the present knowledge on health effects of exposure to radon and its progeny underground, which is used for risk extrapolation to the indoor environment by the BEIR-VI Committee (8) is 6

7 based on cohorts including about 60,000 miners all together. This database will be doubled within the next decades by the study on German Wismut miners. REFERENCES: 1. J.H.Lubin, J.D.Boice, C.Edling, R.W.Hornung, G.H.Howe, E.Kunz, R.A.Kusiak, H.I.Morrison, E.Radford, J.Samet, M.Tirmarche, A.Woodward, YS.Xiang, D.A.Pierce. Radon and Lung Cancer Risk. A joint analysis of 11 underground miners studies. Washington:U.S. Department of Health and Human Services, National Institute of Health. NIH Publ. No :1-136 (1994) 2. S.C.Darby, E.Whitley, G.R.Howe, S.J.Hutchings, R.A.Kusiak, J.H.Lubin, H.I.Morrison, M.Tirmarche, L.Tomasek and S.X.Yao, Radon and cancers other than lung cancer in underground miners: a collaborative analysis of 11 studies; J. Natl. Cancer Inst. 87: (1995) 3. I.Brüske-Hohlfeld, M.Möhner and H.E.Wichmann, Predicted numbers of lung cancer cases in Germany among former uranium miners of the Wismut. Health Phys. 72:3-9 (1997). 4. H.Wesch, T.Wiethege, A.Spiethoff, K.Wegener, K.M.Müller and J. Mehlhorn. German uranium miner study - Historical background and available histopathologic material. Radiat. Res. 152(6):S48-S51 (1999) 5. F.Lehmann, L.Hambeck, KH.Linkert, H.Lutze, H.Meyer, H.Reiber, HK.Renner, A.Reinisch, T.Seifert, F.Wolf. Belastung durch ionisierende Strahlung im Uranerzbergbau der ehemaligen DDR. St. Augustin, Hauptverband der gewerblichen Berufsgenossenschaften (1998). [in German] 6. T.Wiethege, H.Wesch, K.Wegener, K.M.Müller, J.Mehlhorn, A.Spiethoff, D.Schömig, M.Hollstein and H.Bartsch. German uranium miner study - Pathological and molecular genetic findings. Radiat. Res. 152(6): S52-S55 (1999) 7. M.Kreuzer, B.Grosche, A.Brachner, K.Martignoni, M.Schnelzer, H.J.Schopka, I.Brüske-Hohlfeld, H.E.Wichmann and W.Burkart. The German uranium miner cohort study: Feasibility and first results. Radiat. Res. 152 (6):S56-S58 (1999). 8. Committee on Biological Effects of Ionizing Radiation (BEIR VI). Health effects of exposure to radon - BEIR VI. Washington DC, National Academy Press (1998). 7